Mudra

Hardware, Python, Next.js

๐Ÿ‘Œ Next-generation human-computer interaction through biophysical intelligence โ€” our Make-A-Ton 8.0 submission

Cover image for Mudra

Source on GitHub

Mudra (Malayalam: เดฎเตเดฆเตเดฐ) is a multi-purpose wearable input device that combines EMG (muscle activity) and IMU (motion) signals to recognize gestures in real time. By fusing these data streams it achieves more robust and accurate gesture recognition than traditional camera-based systems, which are limited by line-of-sight and lighting. Predictions are streamed instantly to a live, reactive interface via WebSockets.

It was our submission for Make-A-Ton 8.0, conducted by CUSAT.

๐Ÿš€ Key capabilities

  • EMG & IMU fusion โ€” robust gesture recognition without any camera
  • Low-latency streaming โ€” instant WebSocket broadcasting for real-time feedback
  • Privacy-friendly โ€” all interaction derived from biometric and inertial signals
  • Dynamic accumulation โ€” the frontend builds language and geometry in real time from incoming gesture streams
  • Neo-brutalist aesthetic โ€” high-impact raw design language

๐Ÿ—๏ธ Architecture

  • Backend (Python) โ€” UDP receiver โ†’ feature extraction โ†’ ML prediction (Random Forest) โ†’ WebSocket server broadcasting state to clients
  • Frontend (Next.js 15 + React) โ€” smart content accumulation logic, a dynamic layout engine, and webcam-ready AR overlays

Built for the creators of interactive experiences.

← All projects